Data-driven shape sensing of a hypersonic inlet ramp

Abstract

This paper introduces shape sensing using fibre Bragg gratings (FBGs) for hypersonic applications. Accurate shape sensing for geometry sensitive surfaces such as inlet ramps and control surfaces is crucial for monitoring and maintaining vehicle performance, stability, and integrity. Firstly, a simple cantilever plate is examined from which the shape is reconstructed using both analytical and data driven approaches. Following this, surrogate data is generated from a finite element model of a scramjet inlet ramp to train several neural networks. The results show disparity between measured and predicted strains, likely as the result of errors in the model or in the bonding of the FBGs to the ramp. Strain correction factors are obtained from static load testing which improves the predictive capabilities of the network under static loads but reduces accuracy under dynamic loading, indicating the need for an autoregressive component.